Wintersemester

Zeitplan

VU Computer Vision

Zeitplan Vorlesungsteil

Die Vorlesung findet geblockt am Dienstag  (EI 8) von 14:00 bis 16:00, sowie Donnerstag (EI 10) und Freitag (EI 10) von 13:00 – 15:00 c.t. statt, Beginn ist am 12.10.2017. Es folgt eine (vorläufige) Aufstellung der Lehrinhalte:

  • 12.10.2017: Vorbesprechung und Einleitung: Was ist Computer Vision?
    • Organisation: Ziele der Lehrveranstaltung, Ablauf der LVA
    • General introduction
    • History of Computer Vision
    • Why is Vision Hard?
  • 13.10.2017: Structure of Images 1: Image Formation
    • Image Geometry and Formation
    • Types of Projections
  • 17.10.2017: Structure of Images 2: Texture, Scenes, and Context1
    • Properties of Light
    • Light Transport – BRDF
    • Structure in Images: Visual Features, Filtering, Transformations
  • 19.10.2017: Structure of Images 3: Texture, Scenes, and Context 2
    • Linear Image Transform
    • Edges
    • Local Image Analysis
  • 20.10.2017: Local & Multiscale Representations
    • Wavelets
    • Local Image Representations
    • Multiscale Image Represenations
    • Haar Transform
  • 24.10.2017: Image Analysis
    • Interest Points, Corners
    • Image Analysis
    • Scene Emergent Features
  • 7.11.2017: Scenes 1
    • Scene Recognition
    • Bag of Words
    • SIFT
  • 17.11.2017: Scenes 2
    • Clustering
    • Pyramid Matching
    • Support Vector Machine
    • Classification
  • 21.11.2017: Generative Models for Machine Learning
    • Bayes’ rule
    • Density estimation
    • K-means & EM
    • Principal Component Analysis (PCA)
  • 24.11.2017: Depth Perception
    • Monocular Cues to Depth
    • Absolute Monocular Cues to Depth
    • 3D Object Categorization
    • Pose Invariance
  • 28.11.2017: Single View Reconstruction
    • Camera Geometry and Calibration
    • Three-dimensional reconstruction from single and multiple images
    • Single View Metrology
  • 30.11.2017: Absolute 3D Reconstruction
    • Absolute Monocular 3D Reconstruction
    • Absolute Binocular 3D Reconstruction
    • Stereo Geometry
    • Essential Matrix
  • 1.12.2017: Multiview Geometry
    • Correspondence Analysis
    • Multiview Geometry & RANSAC
    • Shape from Motion